Abstract: The aim of this study was to research the applicability of the CERES-Wheat Model and to use the model to predict the winter wheat yield. Based on the parameters calibrated correctly, constant 3a experimental data has been used for evaluation. After input 2007–2008 weather data of DAVIS station, the model predicted that the yield of winter wheat in 2008 was 6 000 kg/hm2. Then CERES-Wheat Model simulated and observed results (including the phonological dates, LAI and the yield)was compared. The final results showed that the deviation of simulated and observed anthesis dates was less than 6 days, the deviation of simulated and observed maturity dates was less than 5 days, simulated LAI and yield were higher than field-observed results, while the deviation of yield prediction was less than 10%. So the applicability of CERES-Wheat Model in Beijing is good and the model can be a useful research tool for yield forecast in agriculture.